Chemical Engineering Journal 321 (2017) 233–244

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Pharmaceuticals removal by activated carbons: Role of morphology on

cyclic thermal regeneration
Susana C.R. Marques a, Jossano M. Marcuzzo b, Mauricio R. Baldan b, Ana S. Mestre a, Ana P. Carvalho a,⇑
a
Centro de Química e Bioquímica, Faculdade de Ciências, Universidade de Lisboa, 1749-016 Lisboa, Portugal
b
Instituto Nacional de Pesquisas Espaciais – INPE, São José dos Campos, SP, Brazil

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Cloth, granular and powder carbons

were tested as adsorbents of
paracetamol and clofibric acid.

Adsorption mechanism of both
compounds is ruled by the carbons’
supermicropores volume.

Higher and quicker removal of the
pollutants are achieved with cloth
and powder samples.

Thermal regeneration cloth and
granular carbons at 400 °C recovered
high paracetamol uptake.

a r t i c l e i n f o a b s t r a c t

Article history: This work aims to evaluate the performance of activated carbons as reusable adsorbents of pharmaceu-
Received 9 January 2017 tical compounds. To achieve this objective, the behaviour of carbons with different morphologies (pow-
Received in revised form 21 March 2017 dered, granular and cloth) in the adsorption of paracetamol and clofibric acid from aqueous solution was
Accepted 22 March 2017
studied; as well as the thermal regeneration of paracetamol saturated activated carbons at 400 and
Available online 23 March 2017
600 °C. For that, the properties of the carbon materials were characterized by N2 and CO2 adsorption,
pHPZC, XPS, TG, XRD and SEM. Kinetic results showed the importance of supermicropore volume for
Keywords:
the diffusion of the probe molecules towards the adsorption active sites, and the negative effect of gran-
Activated carbon
Morphology
ular form which led to significantly lower adsorption rates. Paracetamol adsorption followed a Langmuir
Paracetamol mechanism in almost all cases, whereas clofibric acid adsorption generally occurred through a more com-
Clofibric acid plex mechanism. This behaviour was explained considering the nature of the clofibric acid species pre-
Adsorption from solution sent in solution. This compound was always the most adsorbed molecule, reaching a maximum
Thermal regeneration adsorption capacity of 500 mg dm3 in the case of carbon cloth.
Thermal regeneration proved to be an efficient methodology to recover the porosity of the granular and
cloth paracetamol exhausted activated carbons. After the second regeneration treatment at 400 °C both
activated carbons retained around 57% of their initial paracetamol uptake.
The conjugation of the adsorption and regeneration results pointed out the benefits of the carbon cloth
(in felt form) morphology which, being an easy handling sample, gathered the performance of the best
powdered sample assayed in adsorption experiments, with the behaviour of granular carbon upon
regeneration.
Ó 2017 Elsevier B.V. All rights reserved.

⇑ Corresponding author.
E-mail address: ana.carvalho@fc.ul.pt (A.P. Carvalho).

http://dx.doi.org/10.1016/j.cej.2017.03.101
1385-8947/Ó 2017 Elsevier B.V. All rights reserved.
234 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
1. Introduction
Pharmaceuticals are compounds with biological activity, devel-
oped to promote human health and well-being. Nevertheless, only
a fraction of the pharmaceutical active compound present in, for
example, a painkiller tablet, is metabolized by the organism. There-
fore, a great amount of medicines is excreted, finding their way
into the environment, as it is demonstrated in several monitoring
studies [1–6]. For the scientific community the detection of phar-
maceutical compounds in several sort of water bodies is a long
time recognized problem [7], and the recent inclusion of some
medicines (namely diclofenac) in the Watch List of 2015/495/EU
directive reveals that this issue is now calling the attention of leg-
islative authorities [8].
The present study is focused on paracetamol and clofibric acid
which are among the most detected pharmaceutical compounds
in the aquatic environment. Paracetamol is one of the most con-
sumed analgesic and antipyretic medicines worldwide, being a
major ingredient in numerous cold and flu tablets formulation.
Because of its relatively high solubility in water, it is frequently
detected in the aquatic environment [9–11]. Clofibric acid is the
metabolite and active principle of several blood lipid regulators,
being considered a potential endocrine disruptor; it is nowadays
regarded as one of the most persistent drug residues in the envi-
ronment [12,13]. The first data regarding the detection of this pol-
lutant were reported in the 70’s when it was found in raw and
treated wastewater at concentrations up to 2 mg dm3 [14,15]. In
1993, this chemical was detected in groundwater, in the Berlin
area and, a year later, its presence was confirmed in ground, river
and even in tap water [16].
Activated carbons are versatile adsorbent materials that have
been employed in a large number of industrial processes, including
technologies for the removal of organic pollutants from water (i.e.,
purification of drinking water and wastewater treatment)
[1,17,18]. This thematic has been explored by our research group
in several studies regarding the preparation and use of biomass
derived activated carbons which proved to be efficient adsorbents
for the removal of various pharmaceutical compounds from aque-
ous solutions [19–23].
When activated carbons reach their saturation limit they fail to
adsorb the targeted pollutants. The exhausted materials are, in
many cases, simply disposed in a landfill or incinerated. However,
aiming to a more sustainable society, and to prevent environmen-
tal contaminations, the regeneration of these materials has become
an important issue both from academic and industrial perspec-
tives, as well as from an economic point of view. In this context,
companies like Cabot-Norit provide reactivation of granular acti-
vated carbons through a thermal treatment of the exhausted mate-
rial [24].
Besides thermal regeneration [25–28] other regeneration pro-
cedures have been proposed in the literature, such as, desorption
induced by microwave radiation [29,30], solvent extraction [31],
and by chemical and catalytic decomposition [26], or microbial
processes [32]. The most attractive option is the one that gathers
the best compromise among the cost, the regeneration efficiency
and the number of saturation-regeneration cycles that allows the
recovery of a considerable degree of carbon’s porosity.
Thermal treatment is the most used method for regeneration of
activated carbons due to its simplicity and versatility since, as
mentioned above, it consists in heating the exhausted carbon, usu-
ally under inert atmosphere, to degrade the adsorbate and so
recover the largest possible fraction of the porosity. However, it
must be stressed that the choice of the more adequate operational
conditions to perform regeneration depends on the system formed
by carbon and pollutant(s). In the literature, there are several stud-
ies in this thematic, as for example those developed by Ledesma
et al. [27], and by Sabio et al. [28] both focused on the regeneration
of granular carbons exhausted with p-nitrophenol. Ledesma et al.
[27] promoted five regeneration cycles by heating the exhausted
samples at 900 °C which caused a progressive and accentuated loss
of the textural properties. Sabio et al. [28] extended the study to
more moderated temperatures (300–800 °C) promoting only a sin-
gle regeneration cycle. The results showed that temperatures
lower than 500 °C led to less efficient recoveries of the p-nitrophe-
nol upload. The best result was obtained with 800 °C, allowing the
recovery of 75% of the initial p-nitrophenol uptake.
Considering the regeneration of activated carbons exhausted
with pharmaceutical compounds, fewer studies are available in lit-
erature. Examples are the works developed by Ania and co-workers
over carbons exhausted with salicylic acid which were submitted
to 6 regeneration cycles of microwave assisted treatment [30], or
thermal treatment at 850 °C [25]. In this last case the final sample
retained half of the salicylic acid capacity of the fresh sample. More
recently, Batista et al. [33] reported the regeneration of rapeseed
derived carbons exhausted with caffeine through regeneration at
400, 500 and 600 °C demonstrating that after two regeneration
cycles for 1 h, at the highest temperature, around 95% of the initial
adsorption capacity for caffeine was recovered.
To the best of our knowledge the studies focused on the thermal
regeneration process of carbons are made over only one type of
morphology, generally, powder [25,33,34] or granular [27–29] car-
bons. However, as the regeneration of exhausted activated carbons
is becoming a reality in several processes, and carbons with novel
morphologies have improved performances in industrial processes,
to evaluate the role of the carbons morphology on the feasibility of
the regeneration process is of fundamental importance. So, in the
present study the performance of activated carbons with different
morphologies – granular, powdered and cloth (felt form) – for the
adsorption of two pharmaceutical compounds (clofibric acid and
paracetamol) was evaluated. Consecutive re-uses of thermal regen-
erated paracetamol exhausted carbons were also addressed to
evaluate not only the impact of the heating temperature but also
to shed light on the influence of sample morphology in the regen-
eration process.
2. Materials and methods
2.1. Activated carbons
The lab-made activated carbon felt, herein named Activated
Carbon Cloth, ACC, was prepared from 5.0dtex polyacrylonitrile
(PAN) heavy tow textile fibers following the procedure detailed
described in Ref. 35. Briefly, the textile fibers were submitted to
a two-step air thermal oxidation process in a laboratory scale oven
that operate at temperatures between 200 and 300 °C. The oxi-
dized textile PAN fiber was converted in felt form by standard tex-
tile process. The oxidized PAN fiber in felt form was then
carbonized and activated in the same electrical furnace. In the first
step the material was carbonized under argon flow at 900 °C for
20 min (heating ramp 30 ºC min1). For the activation the argon
gas was shifted to CO2 and the temperature rised at 30 ºC min1
up to 1000 °C, this temperature was maintained during 50 min.
In parallel with this sample three activated carbons commonly
used in wastewater treatment were also assayed:
– NS, powdered carbon commercialized by Norit as SAE SUPER,
with 97% (wt./wt.) of particles with dimensions <0.15 mm.
– VP, powdered carbon V Plus from ChiemiVall, obtained by
steam activation of pine wood. According to the technical report
90% of the particles are smaller than the n° 325 US Standard
Sieve, i.e., smaller than 44 mm.
 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
– GAC, granular carbon prepared by steam activation of coal with
88% (wt./wt.) of particles with dimensions between 0.60 mm
and 2.36 mm. It is commercialized by Norit as GAC830 as being
particularly suitable for thermal regeneration, like it is reported
in the company website [36].
2.2. Characterization of the activated carbons
The nanotexture of the carbon materials was characterized by
N2 and CO2 adsorption at 196 and 0 °C, respectively. The N2
adsorption assays were carried out in an automatic volumetric
apparatus (ASAP 2010 from Micromeritics), and CO2 adsorption
isotherms were obtained in a conventional volumetric installation,
equipped with a pressure sensor MKS-Baratron 310BHS-1000
(0–133 kPa). Before the experiments, the samples (50 mg) were
outgassed for 17 h at 120 °C, under vacuum better than 102 Pa.
The surface chemistry of the carbon cloth sample was charac-
terized determining the pH of the point of zero charge, pHPZC, fol-
lowing the procedure reported by Babić et al. [37]. Solutions of
KNO3 (0.01 mol dm3) with initial pH ranging from 2 to 12 were
prepared mixing different volumes of KOH and HNO3
(0.1 mol dm3) solutions. A volume of 10 cm3 of each of the
KNO3 solutions prepared was added to 50 mg of ACC, in different
vials. After introducing a magnetic stir, the vials were sealed under
N2 atmosphere and stirred at 700 rpm, for 24 h. The pHPZC value of
ACC was obtained from the back-extrapolation of the plateau
defined in the graphic of final vs initial pH values. The surface
groups were also characterized by X-ray photoelectron spec-
troscopy (XPS). The measurements were carried out in a Kratos
Axis Ultra XPS spectrometer using a monochromatic Al-Ka
(1486.5 eV) X-ray radiation. The emitted photoelectrons were
detected using a hemispherical analyzer and 15 lm spatial resolu-
tion. The vacuum system was maintained at approximately 107 Pa
during all the experiments. Survey scans were collected from 0 to
1200 eV with 160 eV pass energy and step size of 1 eV, in order to
identify the elements present on the surface, and a pass energy of
40 eV for high resolution scans on specific atomic peaks
(280–300 eV and 525–543 eV). For calibration purposes C 1s band
(284.6 eV) was taken as internal standard. The deconvolution of
the XPS spectra was made using Casa XPS software considering a
non-linear least square fitting. Before deconvolution a Shirley type
background subtraction was made.
The surface morphology of ACC was analyzed by scanning
electron microscopy in a NovaNanoSEM400 equipment using an
accelerating voltage of 20 kV.
The ash content of the carbons was determined following the
procedure described in ASTM D2866 [38]. The mineral matter
obtained after total carbon matrix consumption was characterized
by X-ray powder diffraction (XRD) using a Pan’ Analytical
PW3050/60X’Pert PRO (h/2h) equipped with X’Pert Data Collector
detector software) using a monochromatized CuKa radiation as
incident beam, 40 kV–30 mA. The diffractograms were obtained
at room temperature by continuous scanning in a 2h range of
15–40° with a step size of 0.017° 2h and a time per step of 20 s.
2.3. Liquid phase adsorption assays
The target molecules selected for the liquid phase assays were
paracetamol (Aldrich, Lot. 535764-326, purity 98%), and clofibric
acid (Alfa Aesar GmbH & Co., KG (Lot. G1266B)). All the solutions
were prepared using ultra-pure water obtained from Milli-Q water
purification system. The solutions were used as prepared, that is,
without pH adjustment, presenting pH values around 5 and 3 units,
for paracetamol and clofibric acid, respectively.
To study the adsorption kinetics, a volume of 20 cm3 of solution
(initial concentration 120 mg dm3) was added to approximately
235
6 mg of adsorbent in glass vials. A magnetic stir bar was intro-
duced, the vials sealed and placed in a thermostatic bath at 30 °C
(Eurotherm 2216 L), stirring at 700 rpm in a multipoint agitation
plate (Variomag Multipoint). Samples were collected between
1 min and 24 h. After filtration, the amount of adsorbate remaining
in solution was determined by UV–Vis spectrophotometry
(Genesys 10S) at its maximum absorbance wavelength (243 nm
for paracetamol and 228 nm for clofibric acid). The adsorbate
uptake was calculated according to:
C0  Ct
qt ¼ V ð1Þ
W
where qt is the amount (mg g1) of compound adsorbed at time t. C0
is the adsorbate initial concentration (mg dm3), Ct is the adsorbate
concentration at time t (mg dm3), V is the volume (dm3) of the
adsorbate solution and W is the weight (g) of dried carbon.
Equilibrium adsorption studies were performed at 30 °C, varying
the adsorbent dose (around 3–6 mg), solution volume (9–30 cm3),
and adsorbate initial concentration (20–180 mg dm3). After
stirring overnight, the concentration of adsorbate in solution at
equilibrium (Ce) was determined and the uptake was calculated,
using Eq. (1).
Prior to use, the carbons were dried overnight in a ventilated
oven at 100 °C, and it must also be mentioned that the assays were
made, at least, in duplicate.
2.3.1. Regeneration treatment assays
Regeneration of cloth (ACC), powder (NS) and granular (GAC)
carbons was performed on the exhausted samples, obtained by
contact with the paracetamol solution (180 mg dm3) using the
proportion of approximately 6 mg of carbon/30 cm3 of solution.
After equilibrium, the solids were recovered by filtration and dried
at 100 °C for 24 h. To perform the thermal regeneration, the sam-
ples were placed in a ceramic boat and heated in a horizontal fur-
nace (Thermolyne, model 21100) at 400 and 600 °C for 1 h, under
N2 flow of 5 cm3 s1 (heating rate of 10 °C min1). After cooled
down to room temperature the samples were weighted and
re-exhausted using the experimental conditions above mentioned.
To establish the regeneration conditions, thermogravimetric
analysis of the exhausted samples was performed on a Setaram
Labsys apparatus, under N2 flow (30 cm3 min1) and at a heating
rate of 15 °C min1 up to 900 °C. Thermogravimetric analysis of
the raw carbon materials and paracetamol were also made in the
same experimental conditions.
The efficiency of each regeneration cycle was quantified by the
regeneration efficiency, RE(%), defined as
qi
REð%Þ ¼  100 ð2Þ
q0
where qi is the adsorption capacity of the regenerated carbon in a
given ith re-use cycle, and q0 the adsorption capacity of the fresh
carbon.
The textural properties of the pristine samples heated under
regeneration conditions, and after the 3rd re-use were assessed
by N2 adsorption data. The fresh carbons submitted to the thermal
treatment will be named according to Carbon/Regeneration
Temperature (°C), and the samples regenerated after the 3rd
re-use cycle will be designated by Carbon exh/Regeneration
Temperature (°C).
To evaluate the possible influence of the carbons GAC and NS
mineral matter on their performance as adsorbents of paracetamol,
the ashes obtained after total consumption of the carbon fraction
were tested as adsorbents. The assays were made at 30 °C, for
24 h, using 6 mg of ashes and 30 cm3 of 180 mg dm3 paracetamol
solution.
236 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244

3. Results and discussion Table 1

Nanotextural properties and pH at the point of zero charge (pHPZC) of
the carbon cloth sample (ACC).
3.1. Characterization of the carbon adsorbents
Parameter Value
The N2 adsorption-desorption isotherms at 196 °C (Fig. 1) ABET (m2 g1) 1136
reveal the different textural characteristics of the samples. In the Vtotala (cm3 g1) 0.76
Vmesob (cm3 g1) 0.03
case of sample ACC, the curve shows a gradual approach to the pla-
Va totalc (cm3 g1) 0.73
teau which is only observed for p/p0 > 0.5. According to IUPAC clas- Va ultrac (cm3 g1) 0.00
sification this is a type I(b) isotherm [39], revealing the presence of Va superc (cm3 g1) 0.73
a broad pore size distribution in the large micropore and small pHPZC 9.0
mesopore range. The configuration of the isotherms obtained with a
N2 volume adsorbed at p/p0 = 0.95.
the commercial carbons suggests also the presence of a micro and b
Vmeso = Vtotal – Va total.
mesopore network, but in these cases the curves are type I + IV iso- c
aS method was applied using as reference the isotherm pre-
therms [39] presenting an upward deviation in the range of high sented in Ref. 41.
relative pressure, more pronounced in the case of carbon NS denot-
ing a more developed mesopore structure. All commercial samples
present small type H4 hysteresis loops [39]. 0.25
From N2 adsorption data, the apparent surface area, ABET, was
determined using BET equation (in the range 0.05 < p/p0 < 0.15)

Differential volume (cm3 g-1 nm-1)

[40]. The microporosity was analyzed applying the aS method tak-
ing as reference the isotherm reported in Ref. 41. With this
method, besides the determination total micropore volume, Va total,
the volumes of ultra (width less than 0.7 nm), Va ultra, and supermi-
cropores (width between 0.7 and 2 nm), Va super, were also discrim-
inated. The results obtained, as well as the meso, Vmeso, and total,
Vtotal, pore volumes are presented in Table 1 and S1 in Supplemen-
tary Information.
The textural parameters reveal that ACC and NS are the adsor-
bents with higher apparent surface areas. Nonetheless, all the
adsorbents have a well-developed micropore network, with Va total
ranging from 0.30 cm3 g1 to 0.73 cm3 g1. On the other hand, with
exception of ACC, all other carbons present also relevant mesopore
volumes, which in the case of carbon NS attains 0.30 cm3 g1, cor-
responding to 44% of the total pore volume of the sample. Regard-
ing microporosity, all the samples have a micropore network
composed mainly by supermicropores, with exception of sample
VP, which presents equal volumes of narrow and wider micropores
(Va ultra = Va super). In sample ACC the microporosity is only com-
posed by supermicropores, and in NS and GAC carbons supermi-
cropore volume corresponds to 95 and 73% of the total
micropore volume, respectively.
Further characterization of the carbons’ micropore structure
was made by CO2 adsorption isotherms from which micropore size
distributions (Fig. 2) were obtained, using the methodology pre-
sented in Ref. 42. These results corroborate the previous discussion
of the textural parameters obtained from the analysis of N2 adsorp-
tion data, showing that the micropore network of the commercial
20
15
nads (mmol g-1)
ACC
0.20
VP
GAC
0.15 NS
0.10
0.05
0.00
0.30 0.50 0.70 0.90 1.10 1.30 1.50 1.70 1.90
Pore width (nm)
Fig. 2. Micropore size distribution of the mentioned samples obtained from the CO2
adsorption data at 0 °C, according to the method described by Pinto et al. [42].
carbons present both narrow and wider micropores, i.e. in all the
cases a broad bimodal distribution was obtained. The maximum
in the region of narrow micropores is located at 0.61 nm, in the
case of carbons VP and around 0.65 nm in the case of GAC and
NS. In any case no pores between 0.96 and 1.13 nm are present.
Sample ACC has a different micropore size distribution, revealing
only the existence of micropores wider than 1.60 nm, which corre-
spond to the highest volume assessed by the CO2 adsorption, in
line with the highest Va total value assessed by N2 adsorption data.
The SEM micrographs of sample ACC reproduced in Fig. 3, show
the structure typical of a carbon cloth, as it was expected for a
material produced from PAN cloth. The yarns observed present a
bean shape, with an average diameter of 21 lm.
Concerning surface chemistry characteristics, the carbons are
basic solids with similar pHPZC values (see Table S1 in Supplemen-
tary Information), allowing to discard the possibility of different
interactions between the adsorbates and the carbon surface as con-
sequence of different surface net charges.
The surface functionalities of samples ACC, NS and GAC were

ACC also characterized by XPS. The survey spectra displayed in Fig. S1

NS in Supplementary Information show that carbon and oxygen are
10 the main elements detected in all the samples. Very small peaks
VP
are also noticed in the region between 100 and 200 eV, assigned
GAC
5 to the presence of Si and Al, in trace amounts. The surface atomic
percentages of carbon and oxygen obtained from the survey spec-
tra are displayed in Table 2 showing that, as expected, carbon is the
0 major component of the samples. Despite these three samples pre-
0.0 0.2 0.4 0.6 0.8 1.0
sent pHPZC values in a relatively narrow range (8.4–9.0), the
p/p0
amount of oxygen quantified by XPS is comprised between 5.2
Fig. 1. Nitrogen adsorption-desorption isotherms at 196 °C of the mentioned and 9.0 atomic%, and the higher oxygen percentage was obtained
samples (closed symbols are desorption points). for the more basic carbon, sample ACC, pointing out that the outer
 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
200 µm
Fig. 3. SEM micrographs of carbon ACC.
Table 2
XPS atomic percentages for samples ACC, NS and GAC.
Atomic %
Sample
C 1s O 1s
ACC 91.0 9.0
NS 93.2 6.8
GAC 94.8 5.2
surface of this carbon particles may present mainly oxygen basic
groups.
All high-resolution C 1s spectra (Fig. 4) have asymmetric tailing
for high binding energy values which is attributed to the intrinsic
asymmetry of the graphitic peak, and to the contribution of oxygen
surface complexes [43]. The spectra were deconvoluted consider-
ing the peaks of the carbon skeleton and of the oxygen functional-
ities usually reported in the literature for this type of materials at
binding energies around: 284.6–285.1 eV, a main peak assigned to
graphitic carbon (sp2); 286.3–287.0 eV, attributed to alcohol or
ether groups (single CAO bond); 287.5–288.1 eV assigned to car-
bonyl and/or quinone groups (C@O); 289.3–290.0 eV, related to
carboxyl or ester groups (COO); and 291.2–292.1 eV, ascribed to
shake-up satellite peaks due to p-p⁄ transitions in aromatic rings
[43–45].
The high-resolution O 1s spectra were fitted considering
three peaks which are commonly reported in the range of
530.4.–530.8 eV (C@O groups), 532.4–533.1 eV (CAO bond in
C-OH and/or CAOAC groups), and 534.8–535.6 eV (quimisorbed
oxygen and/or water) [44].
The deconvolution of the C 1s spectra reveal differences in the
chemistry of the outer surface of the carbon particles. In the spec-
tra of carbons NS and ACC the deconvolution reflected the presence
all the expected five peaks while in the case of the granular mate-
rial the fitting revealed the presence of only four peaks, showing
the absence of C@O related to carbonyl and/or quinone functional-
ities. The deconvolution of the O 1s spectra showed the presence of
the three peaks attributed to C@O, CAO and quimisorbed oxygen
and/or water. In line with the findings of the C 1s data for GAC
sample, the amount of C@O functionalities are considerably smal-
ler than those estimated for the other two samples.
3.2. Liquid phase adsorption
Paracetamol and clofibric acid were the target molecules
selected for the liquid phase assays. The results of the kinetic
assays (Fig. 5) show that, in the experimental conditions used,
the removal reached 50% in the case of paracetamol, and 80% when
clofibric acid was used.
237
10 μm 1 μm
The adsorption kinetic curves obtained for GAC show that, in
comparison with the other carbons, a much slower approach to
the equilibrium is observed, which is in line with the granular mor-
phology of this sample. Actually, regardless the pollutant, equilib-
rium is reached after 6 h of contact time, which was confirmed by
the removal after 24 h.
The adsorption onto the powdered (NS and VP) and cloth
(ACC) carbons resulted in kinetic curves with similar profiles,
revealing a fast adsorption process. The curves show a very
marked decay in the first 5 min, after what the adsorption pro-
ceed more slowly towards equilibrium which was attained after
1 h, in the case of paracetamol, and 2 h when clofibric acid was
tested. Considering the textural characteristics of these samples,
namely the fact that all have high volumes of wider micropores
(see Table S1), this was the expected behaviour according to the
literature. Actually, in previous studies developed in our research
group [21,33,46] it was concluded that, when the samples present
important volumes of supermicropores (i.e. pores with widths
between 0.7 and 2 nm), the presence of a mesopore network does
not have an impact in the diffusion of species with small critical
dimensions, as is the case of paracetamol and clofibric acid. In
fact, according to the values reported in the literature, the critical
dimensions of paracetamol monomer and dimer forms are 0.46
and 0.66 nm, respectively [21], and the critical dimensions of
clofibric acid is 0.70 nm [47], as illustrated in Table S2 of the Sup-
plementary Information. Therefore, no constrains should occur to
the adsorption of these molecules in the supermicropore network
of the samples.
The kinetic data were fitted to the pseudo-first order kinetic
model [48] but the coefficients of determination, R2, were very
unfavorable. The pseudo-second order model [48] leads to better
fittings of the experimental results, with high R2 values and similar
calculated and experimental uptake values (qe calc and qe exp,
respectively) (Table 3).
The values of initial rate, h, corroborate the analysis of the
kinetic curves configuration since they show that the adsorption
of both compounds is a faster process when ACC, NS and VP car-
bons were used. Actually, in the case of these samples the initial
adsorption rate values are two orders of magnitude higher than
those obtained for the granular sample (GAC).
The uptake values at equilibrium show that, with exception of
carbon GAC, clofibric acid is always more efficiently removed than
paracetamol what may be explained considering the solubility of
the compounds, a key factor ruling the adsorption process. The val-
ues reported in the literature [22] are 17390 mg dm3 and
755 mg dm3 for paracetamol and clofibric acid, respectively. Thus,
the lower affinity of clofibric acid towards the solvent, will also
favour the interaction with the carbon surface and, consequently
its removal.
238 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244

Fig. 4. Deconvoluted high-resolution C 1s (left) and O 1s (right) XPS spectra for the mentioned samples. Peaks in C 1s spectra attributed to graphitic carbon sp2 (1), alcohols or
ether groups (2), carbonyl or quinone groups (3), carboxyl or ester groups (4) and shake-up satellite peak due to p-p* transitions in aromatic rings (5) [43–45]. Peaks in O 1s
spectra attributed to C@O groups (1), CAO groups (2) and chemisorbs oxygen and/or water (3) [43–45].

The equilibrium adsorption data for both compounds at 30 °C
(Fig. 6) show that, for all the carbons, clofibric acid is more effi-
ciently adsorbed than paracetamol, most certainly due to its lower
solubility that increases the adsorption potential and consequently
the partition of the clofibric acid towards the adsorbed phase.
Considering the liquid phase isotherms classification for diluted
solution originally proposed by Giles et al. [49] and further simpli-
fied by Lyklema [50], the curves obtained for paracetamol adsorp-
tion can be classified as type L, presenting a relatively well defined
initial rise of the paracetamol uptake followed by a horizontal pla-
teau for equilibrium concentrations, Ce, higher than 40 mg dm3.
In the case of clofibric acid adsorption the isotherm classification is
more difficult, since while the data obtained with carbon ACC can
be classified as a type L curve, the data obtained with carbon VP is a
type F isotherm, due to its less accentuated concavity towards the
concentration axis. The curves obtained with samples NS and GAC
seem to be a mix of L and F type isotherms. Thus, although an
important increase of the clofibric acid uptake is observed up to
Ce values around 40–60 mg dm3 – characteristic of L-type curves
– afterwards the amount of clofibric acid uptake continues to
increase, but more steadily, resembling the curve to a Freundlich
isotherm.
The experimental results were fitted to the linear forms of Lang-
muir [51] and Freundlich [52] models and the resulting parameters
are presented in Table 4. The validation of the adjustments was
made considering the coefficients of determination (R2), and com-
plemented by a chi-square test analysis (v2).
As it was expected from the analysis of the isotherms configu-
ration, paracetamol equilibrium data is better described by the
Langmuir model, i.e. higher R2 and smaller v2 values in comparison
with those obtained for the fitting to the Freundlich equation. Also
in the case of the clofibric acid adsorption, the fitting results cor-
roborate the discussion of the isotherms configuration, showing
that the adsorption process onto ACC and VP carbons is clearly best
 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244 239

(a) consider that paracetamol dimer is the specie present in the

monolayer.
1.0 The rationalization of the clofibric acid results implies a differ-
ent approach because for the majority of the samples the underly-
0.8 ing mechanism of the Freundlich model cannot be disregarded.
GAC Thus, we opted to compare the carbons performance considering
0.6 the amount adsorbed at Ce around 120 mg dm3. The values reveal
C/C0

VP
a trend similar to the one observed for paracetamol monolayer
0.4 NS
capacities. However due to the critical dimension of this molecule
ACC (0.70 nm) we have to consider that the textural parameter ruling
0.2
the clofibric acid adsorption mechanism is the volume of the
0.0 supermicropores, Va super.
0 2 4 6 The different behaviour of clofibric acid, reflected in the iso-
time (h) therms, where a steady approach to the saturation is observed
(F-type isotherm), may be related to the fact that in the experimen-
(b) tal conditions tested, i.e. pH 3, equal fractions of undissociated
1.0 (neutral) and dissociated (negative) forms of clofibric acid are pre-
sent (pKa 3.6). At this pH the carbons have a net positively charged
0.8 surface (pH < pHPZC), so we could expected a high affinity of the
deprotonated specie towards the surface. However, despite no hin-
GAC
0.6 drance to the access of this specie to the pore network is expected,
C/C0

VP as demonstrated in the literature [23], due to its higher solvation

0.4 NS
ACC
0.2
0.0
0 2 4 6
time (h)
Fig. 5. Kinetic results of (a) paracetamol and (b) clofibric acid, at 30 °C for the
carbon samples (6 mg of carbon for 20 cm3 solution with C0 = 120 mg dm3). Lines
show the fitting to pseudo-second order kinetic model.
described by Langmuir and Freundlich equations, respectively. For
the other two carbons, NS and GAC, the fitting results indicate a
good adjustment to both models, since identical R2 and v2 values
are obtained.
Considering the paracetamol monolayer capacity the activated
carbons can be ordered ACC > NS > GAC > VP, which coincides with
the trend of the total micropore volume of the samples. This type of
correlation was observed in previous studies [21,46] and is inter-
preted considering the dimensions of the paracetamol species
(see Table S2 in Supplementary Information) which can access
both the narrow and wider micropores. However, if we take into
account that the micropore volume of the carbon cloth sample,
ACC, is roughly twice the values presented by the other carbons
we could expected a much higher value for paracetamol monolayer
capacity. This result is most probably due to the fact that the
micropore network of this sample presents only pores larger than
1.7 nm (see Fig. 2) which results in an inefficient filling of the
porosity, leading to the waste of some pore volume, even if we
energy the interaction with the carbons surface is shielded. More-
over, a recent study evaluating the influence of water hardness on
the adsorption of clofibric acid onto activated carbons at pH 3 and
8 [47] reported F-type isotherms and higher adsorption capacities
at pH 3. Therefore, we can admit that in the experimental condi-
tions used in the present study the adsorption process must rely
mainly on the interactions of the undissociated clofibric molecules
with the carbon surface which are of dispersive nature and may
allow multilayer adsorption. The steady approach to the saturation
can then be rationalized considering that the amount of species
that can more effectively interact with the carbon is only a fraction
of the total clofibric acid concentration, which is dependent of the
equilibrium between undissociated and dissociated clofibric acid
species. In fact, once the first undissociated species are adsorbed
the equilibrium is shifted towards this specie, leading to a progres-
sive increase of the uptake.
3.3. Regeneration by thermal treatment
The regeneration assays were made considering samples ACC,
NS and GAC and paracetamol as target compound. To gather more
information regarding the characteristics of these materials some
complementary assays were made, as it is presented in the follow-
ing. The results of the ash content determination revealed that car-
bon ACC is an almost ash free material since the mineral matter is
only 0.2% (wt/wt). This result was expected since the XPS basically
indicated C1s and O1s. The ash content is 6.9% and 14.0% (wt/wt)
for GAC and NS, respectively. Further characterization of the NS
and GAC ashes by X-ray diffraction (see Fig. S2 in Supplementary

Table 3
Pseudo-second order parameters for the pharmaceutical compounds adsorption onto the studied carbon samples at 30 °C: k2 is the pseudo-second order rate constant; h is the
initial adsorption rate; t1/2 is the half-life time; qe calc and Ce calc are, respectively, the pharmaceutical compounds uptake and that remaining in solution at equilibrium, both
calculated by the pseudo-second order kinetic model, and qe exp the experimental uptake value.

Paracetamol Clofibric acid

ACC NS VP GAC ACC NS VP GAC
k2  104 (g mg1 min1) 26 40 40 1 12 23 2 2
h (mg g1 min1) 130 179 102 2 137 222 72 3
t1/2 (min) 1.7 1.2 1.7 61 2.4 1.4 2.7 39
qe calc (mg g1) 222 217 169 132 333 313 196 115
qe exp (mg g1) 218 225 170 143 314 304 186 113
Ce calc (mg dm3) 53 52 69 81 20 26 61 86
R2 0.999 0.999 0.999 0.870 0.999 0.999 0.995 0.992
240 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244

Fig. 6. Equilibrium adsorption isotherms of paracetamol and clofibric acid, at 30 °C. Points represent experimental data while lines correspond to the fitting to Langmuir and
Freundlich models. The data concerning the adsorption of paracetamol onto carbons NS and VP were previously reported in Ref. 21 and clofibric acid onto VP in Ref. 47.

Table 4
Langmuir and Freundlich isotherm parameters for the adsorption of paracetamol and clofibric acid adsorption onto the mentioned carbons. Coefficients of determination, R2, and
chi-square test analysis, v2, for all the fittings. Best fitting for each system is indicated in bold.

Paracetamol Clofibric acid

ACC NSa VPa GAC ACC NS VPb GAC
Langmuir equation
qm (mg g1) 322.6 267.7 169.5 250.0 526.3 567.5 321.3 500.0
KL (dm3 mg1) 0.137 0.180 0.097 0.062 0.056 0.028 0.026 0.022
R2 0.996 0.997 0.993 0.969 0.992 0.991 0.785 0.986
v2 6.0 5.3 4.2 18.2 19.7 20.0 31.3 14.5
Freundlich equation
1/n 0.295 0.350 0.239 0.311 0.448 0.659 0.536 0.592
KF (mg11/n (dm3)1/n g1) 80.9 56.4 46.0 49.4 63.8 25.0 19.5 22.7
R2 0.810 0.862 0.868 0.850 0.929 0.976 0.957 0.983
v2 55.6 47.0 9.3 33.8 30.3 14.0 9.3 14.6
a
Results presented in Ref. 21.
b
Results reported in Ref. 47.

Information) showed that in both cases SiO2 is the major crys-
talline structure present, as demonstrated by the most intense
peaks of quartz (ICCD- 33-1161) and cristobalite (ICCD- 39-1425)
at around 26 and 21 °2h, respectively [53]. These findings are in
line with the XPS survey spectra where trace amounts of Si were
detected.
To disclose any possible influence of the mineral matter on the
performance of the carbons as adsorbents of paracetamol from
aqueous phase, the ashes obtained after total consumption of the
GAC and NS carbon fractions were assayed as adsorbents. No
removal was observed after 24 h of contact time, so, even though
these results cannot be straightly compared with the effect of the
mineral matter when distributed in the bulk, they point out that,
by itself the ash has a negligible contribution to the performance
of the carbons as adsorbents.
The experimental conditions for the thermal regeneration stud-
ies of the paracetamol exhausted carbons were defined after ther-
mogravimetry studies. The thermograms (TG) displayed in Fig. 7(a)
show that the raw NS and GAC materials have negligible weight
loss, while carbon cloth ACC losses 10% of its weigh when heat
treated under N2 up to 900 °C. The exhausted samples present sim-
ilar weight loss profiles with weight loss ranging from 14%
(GACexh) to 22% (NSexh). The differential thermogravimetric
(DTG) curves (Fig. 7(b)) show that the most important weight loss
occurs between 300 and 500 °C, and in the case of NS and ACC car-
bons the maximum is centered at 375 °C and 366 °C, respectively,
while for the granular sample it occurs at 411 °C. Comparing the TG
profiles of the exhausted samples with those of paracetamol and
raw carbons (see Fig. 7 and Fig. S3 in Supplementary Information),
it is possible to attribute the weight loss of the exhausted samples
 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
(a)
100
90
Weight (%)
80
70
60
50
100 200 300 400 500 600
Temperature (ºC)
Fig. 7. (a) Thermograms of the raw and exhausted carbons and (b) corresponding DTG curves.
mainly to the removal/degradation of the adsorbed paracetamol,
since the thermogram of paracetamol reveals an almost complete
weight loss in the temperature range between 300 and 400 °C.
The slightly higher temperature needed to remove/degrade parac-
etamol adsorbed onto sample GAC (411 °C versus 366 and 375 °C
for ACC and NS) seems to be linked to the granular morphology
of the material. In fact, in the case of granular particles heat distri-
bution along the particles may be distinct than that occurring in
materials with smaller particles sizes (NS), or fiber-like morpholo-
gies (ACC) where the microporosity is readily available. Specific
interactions could also explain the higher temperature needed in
the case of GAC but equilibrium adsorption data revealed the oppo-
site trend. The adsorption affinity value (KL, Table 4) of GAC for
paracetamol is one order of magnitude lower than the values of
the carbons NS and ACC indicating a weaker interaction between
carbon surface and adsorbed paracetamol.
These results are similar to those reported by Ania et al. [25], for
carbons exhausted with salicylic acid, being the presence of a sin-
gle peak at around 300 °C assigned to the degradation of physically
adsorbed compound. In the same study, the authors also tested
phenol saturated carbons for which the DTG presented two peaks,
being the one at lower temperature assigned to phenol physically
adsorbed, and the one at higher temperature attributed to the
degradation of chemically adsorbed specie. The study developed
by Ledesma et al. [27] with carbons saturated with p-nitrophenol
proved that also in this case the desorption mechanism is a two-
step process. So, according to the results reported in the literature,
the paracetamol adsorption onto the carbons tested in the present
study seems to follow a physisorption mechanism.
The fact that, in the case of GAC sample, the most important
weight loss occurs at higher temperature than that of ACC and
NS carbons may be explained in terms of samples morphology,
porosity and surface chemistry. Actually, in the case of GAC we
can admit that the temperature distribution inside the granular
particles is not uniform. So, for molecules adsorbed in the inner
part of the granules a higher temperature must be applied. On
the other hand, as reported in the literature [34,54], in carbons
with wide pore size distributions, desorption of the adsorbate
may require lower temperatures. So, the behaviour of carbon NS
may be linked to its developed mesopore network, associated to
a high volume of wider micropores (see Table S1). In the case of
sample ACC, we can also admit that desorption is facilitated
because the micropores of this carbon are connected to the surface.
Additionally, the small particle size of sample NS favors the diffu-
sion of the desorbed molecules towards the gas phase in compar-
ison with what is expected to occur in the case of the granular
sample. The surface chemistry may also justify the slightly distinct
surface interactions of GAC with paracetamol since the C 1s XPS
241
(b)
3.0
GACexh
2.5
Deriv. weight loss (%/min)
ACCexh
2.0 NSexh
GACexh 1.5
ACCexh
1.0
NSexh
ACC
GAC 0.5
NS
0.0
700 800 900 100 200 300 400 500 600 700 800 900
Temperature (ºC)
spectra reveals that on the contrary of NS and ACC, in sample
GAC no carbonyl or quinone groups were identified.
According to these results the regeneration experiments were
made at 400 and 600 °C, for 1 h. To evaluate the impact of the heat-
ing regimes on the carbon texture, the samples were submitted to
the experimental conditions chosen for the regeneration assays in
the absence of the paracetamol. The N2 adsorption isotherms (see
Fig. S4 (left) in Supplementary Information) show that the curves
of the treated samples maintain the configuration of the starting
material, and only in for carbon ACC there is a decrease of the
adsorption capacity, most certainly related with the 10% weight
loss observed in ACC thermogravimetric profile (Fig. 7(a)). In the
case of sample GAC the isotherms of the fresh and heated samples
are coincident, and for NS the heating regime causes only a slightly
increase in mesopore volume, especially when a higher tempera-
ture is used.
The textural parameters presented in Table 5 allow us to quan-
tify the impact of the heating treatment on the carbon’s texture.
Regarding the cloth carbon, ACC, the heat treatment leads to an
important loss of apparent surface area and supermicropore vol-
ume. Compared to the initial values, and regardless the tempera-
ture used, the ABET decreases around 25 percentage points, and
Va super about 32 percentage points. On the contrary, the heating
has no significant impact on GAC texture, and a slight positive
effect in the case of the sample NS. The results show that, regard-
less the temperature, the ABET values of the NS treated samples are
quite close to the value of the pristine carbon. The heating treat-
ment only affects the mesopore volume that, in comparison to
the value presented by NS carbon increases a maximum of 27 per-
centage points, when 600 °C were used.
The regeneration efficiency, RE(%), as defined previously in the
experimental section, was assessed in terms of the paracetamol
adsorption capacity retained by the recovered carbons, compared
to the value presented by the fresh sample. The results obtained
(Fig. 8) show that, as expected, the RE(%) values decrease gradually
with the consecutive re-uses of the carbons, being this effect more
accentuated after the 3rd re-use. On the other hand, the influence
of the regeneration temperature on the carbon’s performance
seems to be dependent of the type of carbon used.
The results obtained with ACC sample show that the use of
milder conditions (400 °C) allows to recover somewhat higher
paracetamol adsorption capacities. In these conditions, it is possi-
ble to apply two saturation-regeneration cycles maintaining 56%
of ACC initial paracetamol capacity. Having in mind that, as it
was previously discussed, the heating temperature by itself has a
significant effect on this carbon texture, the decrease of the RE(%)
values may result from the combination of the temperature impact
on the texture and of a partial blocking of the porosity due to
242 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244

Table 5
Nanotextural properties of the fresh activated carbons, samples heated at 400 °C and 600 °C and obtained after the 3rd re-use.

aS methodc
Samples ABET (m2 g1) Vtotala (cm3 g1) Vmesob (cm3 g1)
Va total (cm3 g1) Va ultra (cm3 g1) Va super (cm3 g1)
ACC 1136 0.76 0.03 0.73 0.00 0.73
ACC/400 828 0.51 0.03 0.48 0.00 0.48
ACC/600 851 0.53 0.03 0.50 0.00 0.50
ACCexh/400 435 0.29 0.02 0.27 0.00 0.27
ACCexh/600 449 0.29 0.01 0.28 0.00 0.28
GAC 838 0.47 0.13 0.34 0.09 0.25
GAC/400 848 0.51 0.17 0.34 0.09 0.25
GAC/600 840 0.50 0.16 0.34 0.08 0.26
GACexh/400 166 0.22 0.14 0.08 0.00 0.08
GACexh/600 321 0.27 0.13 0.14 0.00 0.14
NSd 1065 0.70 0.30 0.40 0.02 0.38
NS/400 1083 0.72 0.33 0.39 0.01 0.38
NS/600 1088 0.80 0.41 0.39 0.01 0.38
NSexh/400 185 0.20 0.12 0.08 0.00 0.08
a
N2 volume adsorbed at p/p0 = 0.95.
b
Vmeso = Vtotal – Va total.
c
aS method was applied using as reference the isotherm presented in Ref. 41.
d
Values presented in Ref. 55.

values estimated from the analysis of the N2 isotherms (see

Fig. S4 (right) of Supplementary Information), presented in Table 5,
reveal that the regenerated ACC carbon retains around 40% of the
ABET and Va super values presented by the fresh sample.
Regarding the thermal regeneration of the granular carbon, our
experiments were able to restore 58% of initial paracetamol
adsorption capacity after two saturation-regeneration cycles at
both temperatures. In this case, the decrease of the RE(%) may be
attributed to the progressive obstruction of the porosity, since
the textural characterization of samples heated at 400 and 600 °C
in the absence of paracetamol did not reveal any significant loss
of porosity. The N2 adsorption results of the GAC regenerated sam-
ples (Fig. S4 (right) of Supplementary Information and Table 5)
support this conclusion since there is a significant decrease of ABET
and micropore volume, especially in the case of the sample submit-
ted to heat treatment at 400 °C. The relatively better results in the
3rd and 4th re-use cycles when the treatments were performed at
600 °C may be also linked to the fact that, as demonstrated by
the DTG analysis, paracetamol removal from this carbon requires
a higher temperature than in the case of the other samples.
The less efficient regeneration was achieved with the powdered
sample NS which, according to the results obtained only presents
RE(%) values around 50% after a single regeneration. Further re-
uses of this carbon lead to a much accentuated decrease of parac-
etamol uptake. This behaviour is most likely due to the retention of
paracetamol on the porosity, which as proved by the data pre-
sented in Table 5 for sample NSexh/400, obtained after 3 re-uses,
leads to a drastic decrease of all the textural parameters.

4. Conclusions

The removal of paracetamol and clofibric acid by powdered,

Fig. 8. Regeneration efficiency (RE(%)) of samples ACC, NS and GAC, for consecutive
re-uses after regeneration by thermal treatment at (a) 400 °C and (b) 600 °C for 1 h.
incomplete removal of paracetamol. The loss of ACC porosity is
demonstrated by the textural parameters of the samples obtained
after the 3rd re-use upon regeneration at 400 and 600 °C. The
granular and cloth activated carbons was evaluated by kinetic
and equilibrium assays. The kinetic results showed that, for small
molecules as those tested, the diffusion towards the adsorption
active sites is ruled by the volume of wider micropores. Equilib-
rium data revealed different underlying adsorption mechanisms
for the two target molecules. Paracetamol adsorption follows a
Langmuir mechanism in all the carbons, while in the case of clofib-
ric acid a more complex mechanism seems to be involved, which
was attributed to the nature of the species present in solution. At
high equilibrium concentrations, clofibric acid is always adsorbed
in much larger amounts.
 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
The regeneration through thermal treatment proved to be a
valid strategy to the recovery of the paracetamol exhausted sam-
ples porosity, especially at the lowest temperature tested
(400 °C). The impact of the carbons’ morphology was also demon-
strated, differentiating positively the granular and cloth samples
which could be re-used at least twice keeping adsorption capaci-
ties higher than 50%.
Lastly, the overall results show the potentialities of carbon cloth
to compete with traditional morphologies since it presented
adsorption performances identical, or even better, than powdered
samples, and behaved equally well as granular samples when sub-
mitted to thermal regeneration. Thus aiming to a more sustainable
society where the regeneration of spent carbons must become a
rule, carbon cloth is an excellent option to the presently used gran-
ular morphology.
Acknowledgments
The authors thank the Fundação para a Ciência e Tecnologia
(FCT), Portugal, for financial support to CQB through the project
UID/MULTI/00612/2013, and FAPESP (Brazil) process 12/51087-6
for Activated carbon felt production. SCRM thanks FCT for the
PhD grant, SFRH/BD/91767/2012 and ASM thanks the financial
support of FCT for the Post-doc grant SFRH/BPD/86693/2012. The
authors thank Dr. C.O. Ania for the TG analysis, Quimitejo for pro-
viding carbon VP, and Salmon & Cia for the supply of carbons NS
and GAC.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.cej.2017.03.101.
References
[1] N. Bolong, A.F. Ismail, M.R. Salim, T. Matsuura, A review of the effects of
emerging contaminants in wastewater and options for their removal,
Desalination 239 (2009) 229–246.
[2] J.O. Tijani, O.O. Fatoba, O.O. Babajide, L.F. Petrik, Pharmaceuticals, endocrine
disruptors, personal care products, nanomaterials and perfluorinated
pollutants: a review, Environ. Chem. Lett. 14 (2016) 27–49.
[3] D. Taylor, T. Senac, Human pharmaceutical products in the environment – the
‘‘problem” in perspective, Chemosphere 115 (2014) 95–99.
[4] J. Rivera-Utrilla, M. Sánchez-Polo, M.A. Ferro-Garcia, G. Prados-Joya, R.
Ocampo-Pérez, Pharmaceuticals as emerging contaminants and their
removal from water. A review, Chemosphere 93 (2013) 1268–1287.
[5] C.G. Daughton, Pharmaceuticals Ingredients in Drinking Water: overview of
Occurrence and Significance of Human Exposure, in: Contaminants of
Emerging Concern in the Environment: Ecological and Human Health
Considerations, Am. Chem. Soc., 2010, pp. 9–68.
[6] K.E. Murray, S.M. Thomas, A.A. Bodour, Prioritizing research for trace
pollutants and emerging contaminants in the freshwater environment,
Environ. Pollut. 158 (2010) 3462–3471.
[7] E. Stumm-Zollinger, Biodegradation of steroid hormones, J. Water Pollut. Fed.
37 (1965) 1506–1510.
[8] Directive 2015, 2015/495/EU of the European Parliament and of the Council of
20 March 2015 establishing a watch list of substances for Union-wide
monitoring in the field of water policy pursuant to Directive 2008/105/EC.
Off. J. Eur. Commun. L78/40.
[9] A.P. Carvalho, A.S. Mestre, M. Haro, C.O. Ania, Advanced methods for the
removal of acetaminophen from water, Acetaminophen: Properties, Clinical
Uses and Adverse Effects, Nova Science Publishers Inc, 2012, pp. 57–105.
[10] M.J. Gómez, M.J. Bueno, S. Lacorte, A.R. Fernández-Alba, A. Agüera, Pilot survey
monitoring pharmaceuticals and related compounds in a sewage treatment
plant located on the Mediterranean Coast, Chemosphere 66 (2007) 993–1002.
[11] E. Gracia-Lor, J.V. Sancho, F. Hernández, Multi-class determination of around
50 pharmaceuticals, including 26 antibiotics, in environmental and
wastewater samples by ultra-high performance liquid chromatography-
tandem mass spectrometry, J. Chromatogr. A 1218 (2011) 2264–2275.
[12] S.K. Khetan, T.J. Collins, Human pharmaceuticals in the aquatic environment: a
challenge to green chemistry, Chem. Rev. 107 (2007) 2319–2364.
[13] T. Heberer, Occurrence, fate and removal of pharmaceutical residues in the
aquatic environment: a review of recent research data, Toxicol. Lett. 131
(2002) 5–17.
243
[14] A.W. Garrison, J.D. Pope, F.R. Allen, GC/MS analysis of organic compounds in
domestic wastewaters, in: C.H. Keith (Ed.), Identification and Analysis of
Organic Pollutants in Water, Ann Arbor Science Publishers Inc., 1976, pp. 517–
566.
[15] C. Hignite, D.L. Azamoff, Drugs and drug metabolites as environmental
contaminants: chlorophenoxylsobutyrate and salycilic acid in sewage
treatment effluent, Life Sci. 20 (1977) 337–342.
[16] H.J. Stan, T. Heberer, M. Linkerhägner, Occurrence of clofibric acid in the
aquatic system – is the use in human medical carethe source of the
contamination of surface, ground, and drinking water?, Vom Wasser 83
(1994) 57–68
[17] POSEIDON, Assessment of technologies for the removal of pharmaceuticals
and personal care products in sewage and drinking water facilities to improve
the indirect potable water reuse, Contract No. EVK1-CT-2000-0047, in: http://
www.eu-poseidon.com, 2001–2004.
[18] A.L. Cukierman, Development and environmental applications of activated
carbon cloths, ISRN Chem. Eng. (2013) 1–21.
[19] I. Cabrita, B. Ruiz, A.S. Mestre, I.M. Fonseca, A.P. Carvalho, C.O. Ania, Removal of
an analgesic using activated carbons prepared from urban and industrial
residues, Chem. Eng. J. 163 (2010) 249–255.
[20] A.S. Mestre, A.S. Bexiga, M. Proença, M. Andrade, M.L. Pinto, I. Matos, I.M.
Fonseca, A.P. Carvalho, Activated carbons from sisal waste by chemical
activation with K2CO3 kinetics of paracetamol and ibuprofen removal from
aqueous solution, Bioresour. Technol. 102 (2011) 8253–8260.
[21] M. Galhetas, A.S. Mestre, M.L. Pinto, I. Gulyurtlu, H. Lopes, A.P. Carvalho,
Carbon-based materials prepared from pine gasification residues for
acetaminophen adsorption, Chem. Eng. J. 240 (2014) 344–351.
[22] A.S. Mestre, R.A. Pires, I. Aroso, E.M. Fernandes, M.L. Pinto, R.L. Reis, M.A.
Andrade, J. Pires, S.P. Silva, A.P. Carvalho, Activated carbons prepared from
industrial pre-treated cork: sustainable adsorbents for pharmaceutical
compounds removal, Chem. Eng. J. 253 (2014) 408–417.
[23] A.S. Mestre, M.L. Pinto, J. Pires, J.M. Nogueira, A.P. Carvalho, Effect of solution
pH on the removal of clofibric acid by cork-based activated carbons, Carbon 48
(2010) 972–980.
[24] http://www.cabotcorp.com/solutions/products-plus/activated-carbon/
reactivation-services.
[25] C.O. Ania, J.B. Parra, C. Pevida, A. Arenillas, F. Rubiera, J.J. Pis, Pyrolysis of
activated carbons exhausted with organic compounds, J. Anal. Appl. Pyrol. 74
(2005) 518–524.
[26] R. Berenguer, J.P. Marco-Lozar, C. Quijada, D. Cazorla-Amorós, E. Morallón,
Comparison among chemical, thermal and electrochemical regeneration of
phenol-saturated activated carbon, Energy Fuels 24 (2010) 3366–3372.
[27] B. Ledesma, S. Román, A. Álvarez-Murillo, E. Sabio, J.F. González, Cyclic
adsorption/thermal regeneration of activated carbons, J. Anal. Appl. Pyrolysis
106 (2014) 112–117.
[28] E. Sabio, E. González, J.F. González, C.M. González-García, A. Ramiro, J. Gañan,
Thermal regeneration of activated carbon saturated with p-nitrophenol,
Carbon 42 (2004) 2285–2293.
[29] C.O. Ania, J.A. Menéndez, J.B. Parra, J.J. Pis, Microwave-induced regeneration of
activated carbons polluted with phenol. A comparison with conventional
thermal regeneration, Carbon 42 (2004) 1383–1387.
[30] C.O. Ania, J.B. Parra, J.A. Menéndez, J.J. Pis, Microwave-assisted regeneration of
activated carbons loaded with pharmaceuticals, Water Res. 41 (15) (2007)
3299–3306.
[31] D. Cooney, A. Nagerl, A. Hines, Solvent regeneration of activated carbon, Water
Res. 17 (1983) 403–410.
[32] K. Sodha, S. Panchani, K. Nath, Feasibility study of microbial regeneration of
spent activated carbon sorbed with phenol using mixed bacteria culture,
Indian J. Chem. Technol. 20 (2013) 33–39.
[33] M.K. Batista, A.S. Mestre, I. Matos, I.M. Fonseca, A.P. Carvalho, Biodiesel
production waste as promising biomass precursor of reusable activated
carbons for caffeine removal, RSC Adv. 6 (2016) 45419–45427.
[34] B. Ledesma, S. Román, A. Álvarez-Murillo, E. Sabio, C.M. González-García,
Fundamental study on the thermal regeneration stages of exhausted activated
carbons: kinetics, J. Therm. Anal. Calorim. 115 (2014) 537–543.
[35] J.S. Marcuzzo, C. Otani, H.A. Polidoro, S. Otani, Influence of thermal treatment
on porosity formation on carbon fiber from textile PAN, Mater. Res. 16 (2013)
137–144.
[36] http://www.cabotcorp.com/solutions/products-plus/activated-
carbon/granulated.
[37] B.M. Babić, S.K. Milonjić, M.J. Polovina, B.V. Kaludierović, Point of zero charge
and intrinsic equilibrium constants of activated carbon cloth, Carbon 37
(1999) 477–481.
[38] ASTM D 2866, 2011. Standard test method for total ash content of activated
carbons.
[39] M. Thommes, K. Kaneko, A. Neimark, J. Olivier, F. Rodríguez-Reinoso, J.
Rouquerol, K. Sing, Physisorption of gases, with special reference to the
evaluation of surface area and pore size distribution (IUPAC Technical Report),
Pure Appl. Chem. 87 (2015) 1051–1069.
[40] F. Rouquerol, J. Rouquerol, K. Sing, Adsorption by Powders and Porous Solids –
Principles, Methodology and Applications, Academic Press, San Diego, 1999.
[41] F. Rodríguez-Reinoso, J.M. Martin-Martinez, A standard adsorption isotherm
for the characterization of activated carbons, J. Phys. Chem. 91 (1987) 515–
516.
[42] M.L. Pinto, A.S. Mestre, A.P. Carvalho, J. Pires, Comparison of methods to obtain
micropore size distributions of carbonaceous materials from CO2 adsorption
244 S.C.R. Marques et al. / Chemical Engineering Journal 321 (2017) 233–244
based on the Dubinin-Radushkevich isotherm, Ind. Eng. Chem. Res. 49 (2010)
4726–4730.
[43] C. Moreno-Castilla, M.V. Lopez-Ramon, F. Carrasco-Marin, Changes in surface
chemistry of activated carbons by wet oxidation, Carbon 38 (2000) 1995–
2001.
[44] S. Biniak, G. Szymanski, J. Siedlewski, A. Swiatkowski, The characterization of
activated carbons with oxygen and nitrogen surface groups, Carbon 35 (1997)
1799–1810.
[45] T.J. Bandosz, C.O. Ania, Surface chemistry of activated carbons and its
characterization, in: T.J. Bandosz (Ed.), Activated Carbon Surfaces in
Environmental Remediation, Elsevier, New York, 2006, pp. 159–229.
[46] M. Galhetas, A.S. Mestre, M.L. Pinto, I. Gulyurtlu, H. Lopes, A.P. Carvalho, Chars
from gasification of coal and pine activated with K2CO3: acetaminophen and
caffeine adsorption from aqueous solutions, J. Colloid Interface Sci. 433 (2014)
94–103.
[47] A.S. Mestre, A. Nabiço, P.L. Figueiredo, M.L. Pinto, M.S.C. Santos, I.M. Fonseca,
Enhanced clofibric acid removal by activated carbons: water hardness as a key
parameter, Chem. Eng. J. 286 (2016) 538–548.
[48] Y.-S. Ho, Review of second-order models for adsorption systems, J. Hazard
Mater. 136 (2006) 681–689.
[49] C.H. Giles, D. Smith, A. Huitson, A general treatment and classification of the
solute adsorption isotherm. I. Theoretical, J. Colloid Interface Sci. 47 (1974)
755–765.
[50] J. Lyklema, Fundamentals of Interface and Colloid Science, Vol II - Solid-liquid
interfaces, Academic Press, New York, 1995.
[51] I. Langmuir, The adsorption of gases on plane surfaces of glass, mica and
platinum, J. Am. Chem. Soc. 40 (1918) 1361–1403.
[52] H.M.F. Freundlich, Over the adsorption in solution, J. Phys. Chem. 57 (1906)
385–470.
[53] Powder Diffraction File Alphabetic Index, ICCD, International Center for
Diffraction Data, Swarthmore, PA, 1988.
[54] M.A. Ferro-García, E. Utrera-Hidalgo, J. Rivera-Utrilla, C. Moreno-Castilla,
Regeneration of activated carbons exhausted with chlorophenol, Carbon 31
(1993) 857–863.
[55] A.S. Mestre, S.C.R. Marques, A.P. Carvalho, Effect of cosolvent in the removal of
caffeine by activated carbons, Ind. Eng. Chem. Res. 51 (2012) 9850–9857.